words-and-verbs

name	words-and-verbs
description	Use when adding a new Romanian word, phrase, or verb to CONJI's data files (public/words.json, public/group-1..4.json). Explains entry format, which file a new entry belongs in, and the required duplicate check.

Adding Words and Verbs to CONJI

Data Files Overview

File	Purpose
public/words.json	Vocabulary: nouns, adjectives, adverbs, pronouns, connectors, conjunctions, verb conjugated forms used in context
public/group-1.json	Group I verbs — infinitives ending in -a (excl. -ea)
public/group-2.json	Group II verbs — infinitives ending in -ea
public/group-3.json	Group III verbs — infinitives ending in -e
public/group-4.json	Group IV verbs — infinitives ending in -i / -î
public/group-information.json	Metadata: group/subgroup codes and descriptions

---

words.json — Entry Format

All entries are objects in a flat JSON array. Every entry requires both fields:

{ "value": "romanian word or phrase", "translations": ["english1", "english2"] }

Rules:

• value — the Romanian word/phrase exactly as it appears in learning material. Preserve diacritics (ă, â, î, ș, ț).
• translations — array of English equivalents, ordered from most common to least. Always at least one item.
• Multi-word phrases are allowed (e.g. "pe de altă parte", "de obicei").
• Definite article forms get their own entry when the article changes the learning value (e.g. "carte" + "cartea").
• Do NOT add verbs (bare infinitives or conjugated forms) to words.json — they belong in group files.
• Exception: a verb embedded in a fixed idiom/phrase where the phrase is the vocabulary unit (e.g. "dacă mă gândesc bine") may go in words.json as a phrase.
• Do NOT add the bare infinitive (e.g. "a merge") to words.json — that belongs only in the verb group files.

Examples:

{ "value": "totuși", "translations": ["however", "yet", "still", "nevertheless"] },
{ "value": "în primul rând", "translations": ["first of all", "in the first place", "firstly"] }

---

group-N.json — Verb Entry Format

Each group file has this shape:

{
  "group": "Grupa N - description",
  "verbs": [ ...verb objects... ]
}

Each verb object:

{
  "infinitive": "a [verb]",
  "type": "regular",
  "group": 1,
  "subgroup": 1,
  "infinitive_translated": ["to do something"],
  "conjugations": {
    "prezent":          { "eu": "", "tu": "", "el/ea": "", "noi": "", "voi": "", "ei/ele": "" },
    "perfect_compus":   { "eu": "", "tu": "", "el/ea": "", "noi": "", "voi": "", "ei/ele": "" },
    "viitor_literar":   { "eu": "", "tu": "", "el/ea": "", "noi": "", "voi": "", "ei/ele": "" },
    "viitor_familiar":  { "eu": "", "tu": "", "el/ea": "", "noi": "", "voi": "", "ei/ele": "" },
    "conjunctiv":       { "eu": "", "tu": "", "el/ea": "", "noi": "", "voi": "", "ei/ele": "" },
    "conditional":      { "eu": "", "tu": "", "el/ea": "", "noi": "", "voi": "", "ei/ele": "" },
    "imperativ":        { "tu": "", "voi": "" }
  }
}

Group assignment:

Infinitive ending	File
-a (not -ea)	group-1.json
-ea	group-2.json
-e	group-3.json
-i or -î	group-4.json

Subgroup assignment (from group-information.json):

• Group 1: 1 = -ez pattern · 2 = no -ez · 3 = stem vowel change · 99 = irregular
• Group 2: 1 = -ez/-ează · 2 = stem vowel change · 99 = irregular
• Group 3: 1 = regular · 2 = stem consonant change · 99 = irregular
• Group 4: 1 = -esc pattern · 2 = stem change · 3 = -ăsc pattern · 99 = irregular

type field: "regular" or "irregular"

---

Decision: words.json vs group file

Is it a verb (bare infinitive OR any conjugated/compound form)?
  YES → Is it inside a fixed idiom/phrase (e.g. "dacă mă gândesc bine")?
          YES → words.json (as a phrase)
          NO  → group-N.json only
                (if parent infinitive missing, add it first)
  NO  → words.json

Duplicate check applies to both files. Use the find-duplicates skill before adding anything.

---

Avoiding Duplicates

Before adding any entry, run the find-duplicates skill (node .claude/skills/find-duplicates/scripts/find-duplicates.js) and review any existing value collisions in words.json or infinitive collisions across the group files.

See .claude/skills/find-duplicates/SKILL.md for full usage.

---

Quick Reference — Common Connectors & Discourse Markers

These go in words.json (not verb groups):

{ "value": "în primul rând", "translations": ["first of all", "firstly"] },
{ "value": "în plus", "translations": ["moreover", "in addition", "furthermore", "besides"] },
{ "value": "pe de altă parte", "translations": ["on the other hand"] },
{ "value": "totuși", "translations": ["however", "yet", "still", "nevertheless"] },
{ "value": "de aceea", "translations": ["therefore", "that's why", "for that reason"] },
{ "value": "probabil", "translations": ["probably", "likely"] },
{ "value": "în cazul acesta", "translations": ["in that case", "in this case"] },
{ "value": "dacă mă gândesc bine", "translations": ["come to think of it", "on second thought"] },
{ "value": "de obicei", "translations": ["usually", "normally"] },
{ "value": "apoi", "translations": ["then", "later"] }